Cerebral vascular disease is a leading cause of gait

Transcription

1 1999 Gait Outcome in Ambulatory Hemiparetic Patients After a 4-Week Comprehensive Rehabilitation Program and Prognostic Factors Stefan A. Hesse, MD; Matthias T. Jahnke, MD; Christine M. Bertelt, MD; Carl Schreiner, MD; Daniela Liicke, TA; Karl-Heinz Mauritz, MD Background and Purpose Although gait training is prominent in the rehabilitation of hemiparetic stroke patients, little is known about its outcome and prognostic factors in mildly affected patients. We therefore intended to assess gait in ambulatory stroke patients before and after a 4-week inpatient rehabilitation program based on the neurodevelopmental technique. Methods We measured vertical ground reaction forces by force plates in 148 stroke patients. Variables were stance durations, peak vertical ground reaction forces at heel strike (Fzl) and toe-off (Fz2), loading and deloading rates, time to Fzl, and time to Fz2. The absolute changes for both legs and symmetry outcome were calculated. In addition, we assessed maximal walking speed, walking endurance, stair climbing ability, and the Motricity Index. Results Stance duration, weight acceptance, push-off of both legs, and the stance duration symmetry improved inde- Cerebral vascular disease is a leading cause of gait impairment resulting in long-term disability and handicap. In a study of 60 patients admitted to a general hospital, the authors reported that 25% of all stroke survivors would never be able to ambulate independently, and in 50%, walking speed was still more or less reduced after 3 months. 1 To rehabilitate stroke patients and to improve their gait, physiotherapists apply different treatment techniques, including a functionally oriented traditional approach and other techniques based on neurophysiological models, such as the Bobath neurodevelopmental technique (NDT) and the Brunnstroem, Rood, and proprioceptive neuromuscular facilitation (PNF) concepts. 2 The superiority of a particular treatment approach has not yet been proven in terms of activities of daily living, including general mobility. 38 Gait outcome studies thus far have focused on the walking ability of acute stroke patients admitted to a general hospital and its predictors. 1-9 Little has been done to evaluate the gait outcome and prognostic factors of a comprehensive stroke rehabilitation of ambulatory patients in a later stage of recovery. Received April 5, 1994; final revision received June 25, 1994; accepted June 30, From the Klinik Berlin, Department of Neurological Rehabilitation, Free University of Berlin (Germany). Reprint requests to Stefan Hesse, MD, Klinik Berlin, Kladower Damm 223, Berlin, FRG American Heart Association, Inc. pendent of changes of gait velocity. The symmetry of the ground reaction forces did not improve. Results were even worse for Fzl and the loading rate at the end of treatment. Sex, age, side of hemiparesis, motor strength, stroke interval, and sensory impairment had no influence on the outcome of symmetry. Functional performance did not improve considerably. Conclusions The absolute changes of the ground reaction forces indicated better weight acceptance and push-off of both legs and thus confirmed the efficacy of the neurodevelopmental technique. The symmetry outcome and the functional performance at the end of treatment, however, challenge the efficacy of intensive rehabilitation therapy for 4 weeks in its attempts to restore physiological gait in these mildly affected patients. (Stroke. 1994;25: ) Key Words gait prognosis rehabilitation The purpose of this investigation was to study gait outcome in a large group of mildly affected stroke patients, defined as those who were ambulatory and competent for the most part in the basic activities of daily living. These patients are frequently admitted for a 4-week comprehensive stroke rehabilitation program based on NDT. In a preliminary investigation we studied gait symmetry and functional walking performance, showing no relevant improvement of either variable at the end of treatment for this group of patients. 10 In continuation of this report we decided to investigate a larger number of patients to further elucidate gait outcome after a 4-week rehabilitation program. In addition to assessing gait symmetry, we assessed absolute changes of vertical ground reaction forces, another process-oriented variable of the Bobath technique in which physiotherapists who are trained in NDT strictly control weight acceptance and push-off of both lower limbs. 11 Predicting factors were analyzed to assess which patients were more likely to benefit from the NDT program. We further assessed functional walking abilities relevant to daily living (eg, endurance, stair climbing). Subjects and Methods Subjects The study, which was approved by the local ethical committee, included all hemiparetic patients treated in our department in A total of 156 patients entered the study; 8 patients dropped out, and 148 completed the study. Only the latter were considered for statistical analysis (91 men and 57 women; mean age, 57.1 years [range, 15 to 84 years]). Seventy-

2 2000 Stroke Vol 25, No 10 October 1994 Fz[N] Vertical Force FIG 1. Graph shows force variables analyzed in the vertical plane. Fz1 indicates peak vertical ground reaction force at heel strike; Fz2, peak vertical ground reaction force at toe-off; LR, loading rate; DLR, deloading rate; t1, time to Fz1; and t2, time to Fz2. five patients suffered from a right-sided and 73 from a left-side hemiparesis. The origin of stroke was in all cases a supratentorial lesion, with ischemia in 125 and intracerebral hemorrhage in 23 patients. The mean interval after stroke was days (range, 39 to 962 days), after which time the majority of spontaneous recovery should have occurred. The mean Barthel Index score was 82.3 (range, 75 to 100). Sixty-five patients had impaired proprioception in the affected leg. The patients, who gave informed consent, met the following criteria: (1) 20 m of independent walking without aids; (2) supratentorial lesion, either ischemia or intracerebral hemorrhage; (3) competent for the most part in the basic activities of daily living, with a minimum Barthel Index score of 75; (4) no additional neurological and/or orthopedic deficits impairing ambulation; (5) no heart failure or less than grade 2 on the New York Heart Association scale; and (6) no severe cognitive or communication impairment. Patients were investigated before and after a 4-week inpatient rehabilitation. They received 45 minutes of physiotherapy, based on the Bobath concept, five times per week. The therapy was performed by experienced therapists with additional qualification in the Bobath technique. Additionally, patients were instructed in a self-administered training program for at least 30 minutes daily. Occupational therapy, speech therapy, and neuropsychological training were administered according to individual needs. On average, the patients received 11.6 sessions (range, 0 to 20) of occupational therapy, 2.4 session (range, 0 to 17) of speech therapy, and 4.8 sessions (range, 0 to 20) of neuropsychological training, each lasting 45 minutes. The low amount of speech and neuropsychological therapy reflects that only patients with no severe impairment of communication and cognition were included. Data Acquisition Assessment of Vertical Ground Reaction Forces Vertical ground reaction forces were measured for each leg by means of a gait symmetry system (Kistler AG), which consists of a walkway (8x1.2 m) with two embedded force plates (60x40 cm). When possible, double-step measurements were made. If the stride was not long enough to hit both plates, single-step measurements were performed instead. Patients walked barefoot at their preferred speed. Five single- or double-step measurements were recorded. Measurements were repeated when the step was not representative. Mean curves were then plotted together with seven measured parameters for each leg (Fig 1): foot-floor contact stance time, peak vertical ground reaction forces at heel strike (Fzl) and toe-off (Fz2), and loading and deloading rates (LR and DLR, respectively) in the time from 0% to 80% of Fzl and 0% to 80% of Fz2. The peak forces Fzl and Fz2 were normalized by the body weight (Fzl% and Fz2%, respectively). The temporal occurrence of peak forces, tl and t2, was estimated in terms of percentage of the stance duration. Advantages of the vertical ground reaction forces are minimal intraindividual variability, 12 high accuracy of measurement, simple applicability, 13 and the fact that the peaks and slopes are indicators for weight acceptance and push-off Assessment of Functional Status The following tasks were performed by the subjects: (1) 10-m walk on level ground at maximum gait velocity; (2) walking endurance (self-adopted speed; limit, 600 m); and (3) stair climbing (self-adopted speed with or without handrail; limit, 90 stairs of 16 cm each). Patients wore their preferred shoes and used no aids or orthoses. They were instructed to walk or climb at their preferred speed (except in the 10-m walking test) until they felt unable to continue. Required time, uninterrupted walking distance, and number of steps were recorded. In addition, scores of the Motricity Index of the paretic lower and upper extremities (0 to 100 points) were documented. 15 Statistical Analysis Confidence intervals were calculated for the mean change in each individual functional parameter. Changes were regarded as significant when the corresponding confidence interval did not include zero. Absolute changes (ABS) of each ground reaction force variable were calculated for the affected (aff) and nonaffected (naf) legs before (pre) and after (post) therapy: and X- ABSaff = X-affpost - X-affpre X-ABSnaf=X-nafpost-X-nafpre where X denotes the variable number. Differences (D) of each ground reaction force variable between the affected and nonaffected sides were calculated before and after therapy: and X-Dpre=X-affpre X-nafpre X-Dpost=X-affpost - X-naf post To document a change in symmetry (S) regardless of sign, the following parameters were used: X-S 2 =X-Dpre 2 - X-Dpost 2 A positive sign always indicated an improvement. The X-S 2 values were then retransformed according to the following equation: X-S=X-S 2 /( X-S 2 ) 1/2 For the statistical analysis three Hotelling's T 2 tests were calculated for the sets of X-ABSaff, X-ABSnaf, and X-S (a=.01), and a univariate F test was also performed within these groups of variables (a=.0014). The statistical software SYSTAT was applied. Results Functional Parameters Before and After As shown in Table 1, the functional status of the patients only partially improved. They walked and climbed more quickly, but their endurance (ground level walking and stair climbing) remained virtually un-

3 Hesse et al Gait Study in Hemiparetic Patients 2001 TABLE 1. Functional Gait Parameters Before and After Time for 10-m distance at maximum speed, s Max walking distance (limit, 600 m), m Velocity during endurance walking, m/min Max number of climbed steps (limit, 90) Velocity during step climbing, No. per min Motricity Index in affected leg Motricity Index in affected arm Before ± Cl indicates confidence interval; max, maximum. Values are mean±sd. *CI does not include zero (ie, significant change after therapy). After ± ± ± ± % Cl to -0.21* to to 11.03* to to 10.91* 4.4 to 8.76* 2.07 to 5.59* changed. The improvement in time for the 10-m walking distance, although statistically significant, cannot be regarded as relevant on the basis of the corresponding confidence interval (A does not exceed 1.75). Motor strength of the upper and lower affected limbs measured with the Motricity Index increased significantly. Vertical Ground Reaction Forces As shown in Table 2, for the affected and nonaffected legs, absolute changes could be demonstrated with Hotelling's T 2 tests (P<.01 = a). Univariate F tests revealed significant changes for stance durations, LR, and DLR for both sides. There was a significant change in tl and t2 only for the affected leg. All these significant changes have to be interpreted as an improvement, ie, the stance duration was reduced, Fzl occurred earlier, LR and DLR increased, and Fz2 occurred later on the affected side. Fig 2 shows the vertical force profile of patient 69 (male; age, 67 years; left hemiparesis; time after stroke, 75 days at study admission) before and after therapy. These changes were independent of the improvements in gait velocity, as could be shown by plotting the individual changes in any of the gait variables against the corresponding changes in gait velocity. The R 2 values ranged between 0.4% and 5.0%. Fig 3 displays the dependence of the absolute change of the stance duration of the affected leg on the change of the gait velocity. Gait Symmetry As shown in Table 3, the symmetry parameters changed significantly (P=.002<a) according to the multivariate statistics. Univariate F tests, however, only showed significant improvement for the stance duration, and Fzl% and LR significantly deteriorated. Probability of the improvement of at least four of the seven variables was estimated as 37.8% (95% confidence interval, 28.8% to 47.2%). Plots of X-S 2 against X-Dpre revealed a dependence of both improvement and deterioration of the initial asymmetry, ie, the larger the initial imbalance, the greater the possible change in both directions. Fig 4 shows the dependence of the squared symmetry value (X-S 2 ) of the LR on the absolute difference between the affected and nonaffected legs before therapy. In addition, X-S was plotted against age, initial Motricity Index (weighted average of upper and lower extremities), and stroke interval. No apparent dependencies could be detected, with coefficients of determination (R 2 ) ranging between 0.9% and 6.1%. Furthermore, quantile-quantile plots of X-S (male) versus X-S (female) were made. Similar plots were calculated for the side of hemiparesis (left or right) and TABLE 2. Vertical Ground Reaction Force Parameters of Affected and Nonaffected Legs Before and Absolute Change After Stance duration, s Fz1%, % BW Relative occurrence of Fz1%, ", 4 stance Loading rate, kn/s Fz2%, % BW Absolute Value of Affected Leg Before 0.91 ± ± ± ± ±7.4 Absolute Change of Affected Leg After -0.05±0.15* 2.1 ± ±5.4* 0.7±1.1* -0.04±7.1 Absolute Value of Nonaffected Leg Before 0.99± ± ± ± ±6.7 Absolute Change of Nonaffected Leg After -0.07±0.15* 1.3± ±0.9* 0.68±4.9 Relative occurrence of Fz2%,, 'o stance Deloading rate, kn/s Multivariate probability 67.0± ± ±7.6* * 72.8± ± ± ±0.8* Fz1 % indicates peak vertical ground reaction force at heel strike normalized by body weight (BW); Fz2%, peak vertical ground reaction force at toe-off normalized by BW. Values are mean±sd. Significant change after therapy according to univariate F test with corrected a=.0014.

4 - / 2002 Stroke Vol 25, No 10 October 1994 Fz(N) Fr(N)! :. : : : before. : ; : : after : : : : \^_ ffectei rj\ Tftcted 4-hi- t1 t a la nee Fzi ti LR 12 DLR (kn/a) (kn/s) affected non-affected ; \n / riectet i', Hi stance FH 1 1 LR Fz2 12 DLR (s) (kn/a) (kn/a) J\, :, -, : ew : ; non-affeeti e \I!! ' FIG 2. Graphs show vertical force profile of patient 69 (male; age, 67 years; left hemiparesis; time after stroke, 75 days at study admission) before and after therapy. Fz1 indicates peak vertical ground reaction force at heel strike; BW, body weight; t1, time to Fz1; LR, loading rate; Fz2, peak vertical ground reaction force at toe-off; t2, time to Fz2; and DLR, deloading rate. sensory impairment (present or absent). No marked differences could be seen. Discussion This study intended to investigate the gait outcome of mildly affected hemiparetic patients who were able to walk independently at least 20 m and were competent for the most part in the activities of daily living. The duration of their comprehensive stroke rehabilitation program is normally restricted to 4 weeks, and the rather low frequency of therapies reflects the realistic situation. Functional Walking Performance Although some functional parameters improved significantly (gait and stair climbing velocities, Motricity Index-Foot, Motricity Index-Hand), confidence intervals (particularly for time for 10-m distance at maximum speed) revealed that these changes were not very marked (Table 1), and walking and stair climbing endurance did not improve at all. The minimal changes can probably be explained by the fact that during a Bobath session static exercises (sitting, standing) prevail while the patient is requested to walk slowly and in a controlled manner. Therapists also do not encourage patients to walk by themselves for fear of stereotyped mass synergies." A (gait velocity) FIG 3. Scatterplot shows dependence of absolute change of stance duration of affected leg on change of gait velocity. Vertical Force Parameters Absolute values of the seven ground reaction force parameters before therapy corresponded with those in previous reports: The stance duration of the affected leg was shorter, 16 and the push-off was less pronounced on the paretic side Fzl% and LR were larger on the affected side. 10 ' 12 Global changes of the vertical ground reaction forces revealed significant improvements in five variables on the affected side and three on the nonaffected side. Changes in gait velocity did not predict these parameter changes. Thus, gait kinetics serve as an independent measure of therapeutic effects in this group of patients. This does not exclude the existence of a relation between force parameters and gait velocity for each individual subject, as in the case of healthy subjects. 17 The reduced stance duration for both lower limbs is therefore not exclusively explained by an increase of walking speed, and it does not change in direct proportion to the cycle duration. The double support times tend to increase with the impairment in hemiplegic gait. 16 It may therefore be assumed that the double support times were reduced more than the cycle times with the improvement of gait disability. The fact that gait improved is further supported by a significant increase of the variables, indicating weight acceptance and push-off of both lower limbs. Therefore, patients showed changes that are in accordance with the intentions of the NDT, ie, an improved weight acceptance and push-off. 11 Comparable data are not available, and control studies had different scopes. 38 Symmetry of Ground Reaction Force Parameters Stance symmetry improved, which, together with a reduction of stance durations, confirmed the objectives of NDT. Symmetry of the ground reaction forces, however, did not improve. On the contrary, Fzl% and LR both deteriorated significantly for this group of patients. These findings refer to mean values and do not exclude improvements in individual patients. If it is assumed that a clinically relevant change requires improvement of at least four of seven symmetry parameters, then the probability of improvement is less than 50%. These findings question the achievement of a main

5 Hesse et al Gait Study in Hemiparetic Patients 2003 TABLE 3. Differences Between Affected and Nonaffected Legs Before and Change of Symmetry After Difference Between Affected and Nonaffected Change of Legs Before Symmetry Symmetry of stance durations Symmetry of Fz1 % Symmetry of relative occurrences of Fz1 % Symmetry of loading rates Symmetry of Fz2% Symmetry of relative occurrences of Fz2% Symmetry of deloading rates Multivariate probability -0.09± ± ± ± ± ± ± ±0.13* -0.65±8.64* 0.40± ±0.96* -0.08± ± ±0.79 Fz1% indicates peak vertical ground reaction force at heel strike normalized by body weight; Fz2%, peak vertical ground reaction force at toe-off normalized by body weight. Values are mean±sd. *Significant change after therapy according to univariate F test with corrected a=.o014. goal of NDT: the restoration of gait symmetry as a key element of physiological gait, at least in mildly affected patients. To channel resources, it is desirable to find predictors that patients are most likely to profit from because therapy is both time- and cost-intensive. In a report on the ability to walk 6 months after a stroke, age, visual neglect, and leg power (Medical Research Council grades) best predicted outcome among 113 subjects. 9 We did not find any relevant influence of sex, age, side of hemiparesis, motor strength, stroke interval, or sensory impairment on the outcome of symmetry in these mildly affected patients. The initial state of asymmetry allowed the prediction of the absolute amount of change but not the direction of change (improvement or deterioration). Thus, there were no criteria to predict which of these mildly affected patients might benefit more from the comprehensive stroke rehabilitation program with regard to gait symmetry. Contrary to the literature on general rehabilitation, elder patients or patients after the period of spontaneous recovery did not profit less from NDT. 18 Sensory impairment did not influence the symmetry outcome, which is in agreement with existing reports on its poor LRaf-LRnrf (initial) FIG 4. Scatterplot shows dependence of squared symmetry value (X-S 2 ) of loading rate (LR) on absolute difference between affected (af) and nonaffected (naf) legs before therapy. Parabolas were fitted separately for positive (improvement) and negative (deterioration) changes in symmetry. predictive value regarding activities of daily living and the weak correlation between sensation and walking performance, upright stability, and other aspects of motor control Conclusions In conclusion, stance duration, weight acceptance, push-off of both legs, and stance-duration symmetry improved significantly during the 4-week NDT rehabilitation program. However, the symmetry of ground reaction forces did not improve, although gait imbalance therapy is a primary component of NDT. The functional performance of the patients did not improve considerably. These results thus challenge the efficacy of intensive rehabilitation therapy in its attempts to restore physiological gait in minimally affected hemiparetic patients. The study does not exclude possible improvements in other aspects of motor behavior, such as reduction of spasticity and retraining of upright posture, that are within the scope of NDT. Further conclusions regarding the rehabilitation of mildly affected hemiparetic patients should be drawn with great caution. On the other hand, the patients might have reached a certain level of functional ability as a result of adaptive mechanisms, so that no substantial further improvements of gait could be expected. Acknowledgments This study was supported in part by a grant from the Bundesversicherungsanstalt fur Angestellte and from the Kuratorium ZNS. References 1. Wade DT, Wood VA, Heller A, Maggs J, Langton Hewer R. Walking after stroke: measurement and recovery over the first three months. Scand J Rehabil Med. 1987;19: Dewald JPA. Sensorimotor neurophysiology and the basis of neurofacilitation therapeutic techniques. In: Brandstater ME, Basmajian JV, eds. Stroke Rehabilitation. Baltimore, Md: Williams & Wilkins Co; 1987: Logigian MK, Samuels MA, Falconer J, Zagar R. Ginical exercise trial for stroke patients. Arch Phys Med Rehabil. 1983;64: Dickstein R, Hochermann S, Pillar T, Schaham R. Three exercise therapy approaches. Phys Ther. 1986;66: Lord JP, Hall K. Neuromuscular reeducation versus traditional programs for stroke rehabilitation. Arch Phys Med Rehabil. 1986; 67:89-91.

6 2004 Stroke Vol 25, No 10 October Basmajian JV, Gowland CA, Finlayson MA, Hall AL, Swanson LR, Stanford PW, Trotter JE, Brandstater ME. Stroke treatment: comparison of integrated behavioral physical therapy vs. traditional physical therapy programs. Arch Phys Med Rehabil. 1987;68: Jongbloed L, Stacey S, Brighton C. Stroke rehabilitation: sensorimotor integrative treatment versus functional treatment. Am J Occup Ther. 1989;43: Wagenaar RC, Meijer OG, van Wieringen PC, Kuik DJ, Hazenberg GJ, Lindenboom J, Wichers F, Rijswijk H. The functional recovery of stroke: a comparison between neurodevelopmental treatment and the Bruunstrom method. Scand J Rehabil Med. 199O;22:l Friedman PJ. Gait recovery after hemiplegic stroke. Inl Disabil Stud. 1991;12: Hesse SA, Jahnke MT, Schreiner C, Mauritz KH. Gait symmetry and functional walking performance in hemiparetic patients prior to and after a 4-week rehabilitation programme. Gait & Posture. 1993;1: Davies PM. Steps to Follow. New York, NY: Springer-Verlag, Carlsoo S, Dahllof AG, Holm J. Kinetic analysis of the gait in patients with hemiparesis and in patients with intermittent claudication. Scand J Rehabil Med. 1974;6: Stiissi E, Debrunner HU. Parameteranalyse des menschlichen Ganges. Biomed Tech (Berlin). 1980;25: Gaeys R. The analysis of ground reaction forces in pathological gait, lnt Orthop. 1983;7: Demeurisse G, Demol O, Robaye E. Motor evaluation in vascular hemiplegia. Eur Neurol. 1980;19: Brandstater ME, De Bruin H, Gowland C, Clark BM. Hemiplegic gait: analysis of temporal variables. Arch Phys Med Rehabil. 1983; 64: Nilsson J, Thorstsensson A. Ground reaction forces at different speeds of human walking and running. Ada Physiol Scand. 1989; 136: Jongbloed L. Prediction of function after stroke: a critical review. Stroke. 1986;17: Feigenson JS, McCarthy ML, Meese PD, Feigenson WD, Greenberg SD, Rubin E, McDowell FH. Stroke rehabilitation, I: factors predicting outcome and length of stay: an overview. NY State J Med. 1977;77: Dettmann MA, Under MT, Sepic SB. Relationships among walking performance, postural stability, and functional assessments of the hemiplegic patient. Am J Phys Med. 1987;66:77-90.